Toe Tapping Based Falling Risk Evaluation for Patients With Parkinson's Disease Using Monitoring Insoles

Typical symptoms of Parkinson's disease (PD) include freezing of gait, shuffling steps, and stooped posture, which introduce high risks of falling. Existing studies characterize and analyze PD symptoms through walking using wearable sensors. In addition, toe tapping is often used in clinical motor examination representing rotational agility in lower extremities, which is a critical factor in PD motor functionality and relates to falling. In this letter, we investigate the time and acceleration feature differences between PD patients and their age-matched healthy control (HC) subjects through alternating and synchronized toe tapping. Monitoring insoles are used to collect acceleration data with ten PD patients and nine HC subjects. In total, 87 time-axis and 1152 acceleration-axis features are extracted and analyzed between the two groups. The results suggest acceleration-axis features are with more significance in differing PD patients and HC subjects than time-axis features. The results show toe tapping motion can better distinguish PD patients and HC with an accuracy of 87.65% using the acceleration-axis feature set compared to 57.07% with time-axis feature set. To help PD patient understand their falling, falling risks are evaluated in scores using linear regression, with (error ± standard deviation) = (0.97 ± 0.85) to be compatible to fall risk questionnaire developed by the U.S. Centers for Disease Control and Prevention. Our method and the experiment can be easily transformed into a mobile game for daily life use and provides falling risk scores to PD patients without occupying spaces.